Is CVE-2026-77931 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-77931 being exploited in the wild.

How severe is CVE-2026-77931?

High severity, with a CVSS score of 8.8, published 15 October 2026.

Which products and versions are affected?

Jxxghp MoviePilot through 3.1.5.

How is CVE-2026-77931 attacked?

It is attacked over the network, in a hosted service.

Am I affected by CVE-2026-77931?

Check whether you run Jxxghp MoviePilot at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2026-77931: MoviePilot command injection via prompt injection in execute_command

CVE-2026-77931 · Severity: high · CVSS 8.8 · Published 2026-10-15

MoviePilot is an AI-assisted media management application that offers administrative users an always-available command execution tool. Attackers can embed hidden instructions in web pages, torrent metadata, or media information that the application reads; when an administrator asks the assistant to process this attacker-controlled content, those instructions can trick the AI into executing arbitrary operating system commands on the host with the MoviePilot process privileges, potentially including root access. No direct access to the instance or authentication bypass is required—only that an admin user analyzes attacker-controlled content.

Technologies: Jxxghp MoviePilot. Vendors: Jxxghp.

Quick answers

Is CVE-2026-77931 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-77931 being exploited in the wild.
How severe is CVE-2026-77931?: High severity, with a CVSS score of 8.8, published 15 October 2026.
Which products and versions are affected?: Jxxghp MoviePilot through 3.1.5.
Is there a fix for CVE-2026-77931?: No fixed version is recorded yet. Until patched, restrict administrative users from processing untrusted content (web pages, torrent metadata, or media information from untrusted sources) via the AI assistant, or disable the execute_command tool and other dangerous command-execution features via configuration if the application supports it. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
How is CVE-2026-77931 attacked?: It is attacked over the network, in a hosted service.
Am I affected by CVE-2026-77931?: Check whether you run Jxxghp MoviePilot at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

How to fix

Until patched, restrict administrative users from processing untrusted content (web pages, torrent metadata, or media information from untrusted sources) via the AI assistant, or disable the execute_command tool and other dangerous command-execution features via configuration if the application supports it.

Technical details

An indirect prompt injection vulnerability in MoviePilot's AI agent allows remote code execution when untrusted content (from browse_webpage, search_web, torrent, or media metadata tools) contains instructions targeting the always-available ExecuteCommandTool. The ExecuteCommandTool accepts an LLM-settable confirm_dangerous flag that bypasses the command safety denylist, and runs commands directly via asyncio.create_subprocess_shell with the MoviePilot process's privileges (often root). When an administrator's session processes attacker-controlled content, the injected instructions cause the model to invoke execute_command, satisfying the require_admin permission check already met by the admin's session context.

Affected products

jxxghp MoviePilot through 3.1.5

Timeline

2026-10-15: disclosed

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.